Additive diversity partitioning (α, β, and γ) is commonly used to study the distribution of species-level diversity across spatial scales. Here, we first investigate whether published studies of additive diversity partitioning show signs of difficulty attaining species-level resolution due to inherent limitations with morphological identifications. Second, we present a DNA barcoding approach to delineate specimens of stream caddisfly larvae (order Trichoptera) and consider the importance of taxonomic resolution on classical (additive) measures of beta (β) diversity. Caddisfly larvae were sampled using a hierarchical spatial design in two regions (subarctic Churchill, Manitoba, Canada; temperate Pennsylvania, USA) and then additively partitioned according to Barcode Index Numbers (molecular clusters that serve as a proxy for species), genus, and family levels; diversity components were expressed as proportional species turnover. We screened 114 articles of additive diversity partitioning and found that a third reported difficulties with achieving species-level identifications, with a clear taxonomic tendency towards challenges identifying invertebrate taxa. Regarding our own study, caddisfly BINs appeared to show greater subregional turnover (e.g., proportional additive β) compared to genus or family levels. Diversity component studies failing to achieve species resolution due to morphological identifications may therefore be underestimating diversity turnover at larger spatial scales.